Information for "Unveiling the hidden bride: deep annotation for mapping and migrating legacy data to the Semantic Web"

The success of the Semantic Web crucially depends on the easy creation, integration, and use of semantic data. For this purpose, we consider an integration scenario that defies core assumptions of current metadata construction methods. We describe a framework of metadata creation where Web pages are generated from a database and the database owner is cooperatively participating in the Semantic Web. This leads us to the deep annotation of the database—directly by annotation of the logical database schema or indirectly by annotation of the Web presentation generated from the database contents. From this annotation, one may execute data mapping and/or migration steps, and thus prepare the data for use in the Semantic Web. We consider deep annotation as particularly valid because: (i) dynamic Web pages generated from databases outnumber static Web pages, (ii) deep annotation may be a very intuitive way to create semantic data from a database, and (iii) data from databases should remain where it can be handled most efficiently—in its databases. Interested users can then query this data directly or choose to materialize the data as RDF files.

For the future, there is a long list of open issues concerning deep annotation—from the more mundane, though important, ones (top) to far-reaching ones (bottom):

(1) Granularity: So far we have only considered atomic database fields. For instance, one may find a string “Proceedings of the Eleventh International World Wide Web Conference, WWW2002, Honolulu, Hawaii, USA, 7–11 May 2002.” as the title of a book whereas one might rather be interested in separating this field into title, location, and date. (2) Automatic derivation of server-side Web page markup: A content management system like Zope could provide the means for automatically deriving server-side Web page markup for deep annotation. Thus, the database provider could be freed from any workload, while allowing for participation in the Semantic Web. Some steps in this direction are currently being pursued in the KAON CMS, which is based on Zope. (3) Other information structures: For now, we have built our deep annotation process on SQL and relational databases. Future schemes could exploit Xquery or an ontology-based query language. (4) Interlinkage: In the future deep annotations may even link to each other, creating a dynamic interconnected Semantic Web that allows translation between different servers. (5) Opening the possibility to directly query the database, certainly creates problems such as new possibilities for denial of service attacks. In fact, queries, e.g. ones that involve too many joins over large tables, may prove hazardous. Nevertheless, we see this rather as a challenge to be solved by clever schemes for CPU processing time (with the possibility that queries are not answered because the time allotted for one query to one user is up) than for a complete “no go.”

We believe that these options make deep annotation a rather intriguing scheme on which a considerable part of the Semantic Web might be built.